From beards to bytes: How AI is empowering small business owners in Kenya

"The overall idea was how can we democratize access to AI and the benefits of AI?" says Mutembi Kariuki, co-founder and CEO of Fastagger, the Nairobi startup behind the Auni app. The highest impact is in micro, small and medium enterprises, but "a lot of places, especially in emerging markets, have challenges with connectivity and the kind of devices that they have. We realized that you could actually run AI models on their four-gigabyte smartphones and enable them to have the AI equivalent of an MBA intern on the phone that can analyze their mobile money transactions."

Kariuki and co-founder Jude Mwenda had founded other startups, but they realized that in addition to a lack of connectivity and the latest technology, Africa also lacked data. So, together with another data scientist, Stephanie Njerenga, in 2019 they founded Fastagger, to speed up the creation of data sets for AI by hiring data experts to manually tag data - hence the name, combining "fast" and "tagger."

During the Covid-19 pandemic, a Kenyan business association asked the Fastagger founders to conduct webinars to help struggling businesses stay afloat. That opened their eyes to how owners of businesses with just a few employees did most of their transactions on their phones, and sometimes not even smartphones, without internet service all the time and "on top of that, it's expensive," Kariuki says.

"We realized we had to build something that even if it was sitting on their own personal phone, then so that we can read the SMS for the mobile money transactions and be able to give them that insight, it would have to work offline and on device. When we say 'on device' we mean that you can literally use AI without having to have internet," he says.

"Our bet has been that the African market doesn't have money for GPU compute," says Mwenda, who is Fastagger's chief technology officer. "So, we compress the model, make it smaller and optimize for low-cost constraints."